Gregory Clark Uses Computer over Phone, Predicts "Economic Redundancy" of Working Class

Gregory Clark's basic assumption would seem to be that some people are born idiots. His argument in this Washington Post op-ed goes something like this: real wages of idiots have not increased because the demand for idiot labor has fallen due to the rise of the machines. Soon, the machines will be able to do anything an idiot can do for less than it costs to pay idiot subsistence wages. So idiots will be left “socially needy but economically redundant.” What will we do?! “There is only one answer,” Clark says. “You tax the winners — those with the still uniquely human skills, and those owning the capital and land — to provide for the losers.”

What to say? Tim Worstall says a good deal of it. This diagnosis by Bruce Wilder in Mark Thoma's discussion thread seems to me in about the right neighborhood:

[In his book Farewll to Alms] Clark could find few institutional differences between 12th century England and 20th century Britain. In his mind, Henry II laid down the law of equity securing property rights, and nothing else much mattered. If anything, he regards the 12th century, with its low tax rates, as much more amenable to economic growth and innovation than the burdensome 20th century welfare state. So, I'm not surprised that he doesn't credit institutional changes for changes in income distribution over the last three decades.

So, the shape of his fears that machines will soon displace the near-cretins serving him hamburgers at McDonald's form a unitary theme. He sees his class burdened by taxes to support the no-account lower classes, who are even more useless now than in centuries past, but, perhaps, can reconcile himself to it as his paternalistic obligation.

Here is what I said some time ago at Free Exchange about Clark's baseless truculence toward institutional explanations in economics.

Here's what I think about Clark's op-ed.

First, technological innovation over the past two centuries has been incredibly rapid, and workers have been repeatedly displaced by technology only to move on to different kinds of jobs. Why hasn't technological change so far created much higher rates of unemployment? Does Clark think this is a historical fluke? Why does he think this pattern is about to be broken? Why does he think technological change is finally reaching a tipping point? His failure to address this obvious point at all is glaring. Is this whole conjecture really built on his experience with an automated phone call to United Airlines?

Second, I think that Clark wrongly accepts that real wages toward the bottom of the distribution have not risen. This is, to my opinion, an artifact of mistaken measurement techniques. See Broda and Weinstein. There is no reason to believe that the market forces which have improved standards of living for the poor will not continue to do so. Indeed, Clark's assumptions about the efficiency gains from future technology provide us reason to think the real prices of many goods will continue to decline.

Third, insofar as wages have stagnated toward the bottom, a decline in hours worked for low-skilled workers explains a good deal of it. Doesn't this show Clark is right?! No. It shows that badly structured welfare policy has provided an incentive for many low-skilled workers to work fewer hours in order to qualify for transfer payments. Because experience (hours worked at a task) is a main determinant of skill level, and skill level is a main determinant of wage levels, an incentive to reduce hours worked is an incentive to remain at a lower level of skill and thus wages. See Deere and Welch ($$$).

Fourth, Clark's theory of blood-born idiocy leads him to conclude that there are little or no gains to be had from improvements in education. Here's what he says:

Others see education as a way out of this dystopia. The root problem is, after all, the widening of the income gap between the skilled and the unskilled. Can expanded education give the poorest the tools to resist the march of the machines? I'm skeptical. Already, much of the supposed improvement in high school and college graduation rates has come by asking less of graduates. We can certainly arrange to have everyone “graduate” from high school, but whether they will have the skills needed to make it is doubtful.

This is maddeningly dense. Clark apparently believes the only way to make graduation rates go up is to devalue diplomas by giving them to irremediable idiots. That is to say, idiots are idiots and education can't do anything about that. A more plausible view is that so many young people graduate high school (or don't) with such poor abilities because the American public education system has failed disastrously to provide a minimally acceptable level of training to children who grow up in poor, predominately minority neighborhoods. The best explanation for this failure is an institutional explanation. The political forces in control of public schools in low-income neighborhoods have strong incentives to resist almost every potentially effective reform. There are no competitive markets in educational services for low-income families because such markets are, in effect, against the law. Were low-income families to have access to a competitive market in educational services, there is every reason to believe the quality of training would rise, the real level of ability of high-school graduates would rise, and the portion of high-school graduates prepared to benefit from higher education would rise.

The fundamental bone of contention here is over the fixity or flexibility of the human capacity to gain and improve economically relevant skills. Here is Cato Unbound's issue on IQ.

Fifth, the piece is short-sighted. If robots can crowd out all low-skilled workers, there is no reason they cannot also crowd out all high-skilled workers. See Hanson. Would this be bad? Growth would proceed so rapidly that the returns to even small amounts of capital should be outrageously high. The gap will be between those with income from capital gains and those with none. To prevent this, some version of Clark's recommendation might be desirable. I'd recommend Charles Murray's scheme for replacing the United States' social insurance apparatus with basic income grants and mandatory retirement and medical savings accounts. In a world of doubling-every-fifteen-minutes Hansonian robot growth, the portion of GDP necessary to fund universal grants sufficient to ensure a modestly lavish level of consumption would be so trifling that no one would even notice. For now, we should try to hasten the arrival of this post-human economy, in which case we should try to optimize incentives to innovation and growth. Higher taxes and higher levels of welfare spending is about the opposite of that.